MAINT: Fix packaging metadata and align the dependency files

README.md

@@ -15,10 +15,13 @@ estimating asset prices:
 
 -   **Panel models**:
     -   Fixed effects (maximum two-way)
+    -   Random effects
     -   First difference regression
     -   Between estimator for panel data
     -   Pooled regression for panel data
     -   Fama-MacBeth estimation of panel models
+    -   Specification tests: the Hausman test of random effects against
+        fixed effects and the Pesaran CD test of cross-sectional dependence
 
 -   **High-dimensional Regression**:
     -   Absorbing Least Squares
@@ -28,6 +31,9 @@ estimating asset prices:
     -   Limited Information Maximum Likelihood
     -   k-class Estimators
     -   Generalized Method of Moments, also with continuously updating
+    -   Jackknife IV (JIVE)
+    -   Specification tests: the Cragg-Donald and Kleibergen-Paap rank tests
+        of the first stage (underidentification)
 
 -   **Factor Asset Pricing Models**:
     -   2- and 3-step estimation
@@ -89,35 +95,46 @@ mod = IV2SLS.from_formula('np.log(wage) ~ 1 + exper + exper ** 2 + [educ ~ mothe
 The expressions in the `[ ]` indicate endogenous regressors (before `~`)
 and the instruments.
 
-## Installing
+## Installation
 
 The latest release can be installed using pip
 
 ```bash
 pip install linearmodels
 ```
 
-The main branch can be installed by cloning the repo and running setup
+Binary wheels are available for Windows, macOS and Linux. Without a wheel, pip
+builds linearmodels from source, which requires a C compiler.
+
+The main branch can be installed by cloning the repo and building it with pip.
+The build uses [meson-python](https://mesonbuild.com/meson-python/), and pip
+installs the other build requirements (Cython, NumPy and setuptools_scm)
+automatically.
 
 ```bash
 git clone https://github.com/bashtage/linearmodels
 cd linearmodels
 pip install .
 ```
 
+To work on linearmodels, install the requirements and then build an editable
+install
+
+```bash
+python -m pip install -r requirements.txt -r requirements-dev.txt
+python -m pip install -e . --no-build-isolation
+```
+
 ## Documentation
 
 [Stable Documentation](https://bashtage.github.io/linearmodels/) is
-built on every tagged version using
-[doctr](https://github.com/drdoctr/doctr).
+built by [GitHub Actions](https://github.com/bashtage/linearmodels/blob/main/.github/workflows/generate-documentation.yml)
+when a release is published.
 [Development Documentation](https://bashtage.github.io/linearmodels/devel)
-is automatically built on every successful build of main.
+is built by the same workflow on every push to main.
 
 ## Plan and status
 
-Should eventually add some useful linear model estimators such as panel
-regression. Currently only the single variable IV estimators are polished.
-
 -   Linear Instrumental variable estimation - **complete**
 -   Linear Panel model estimation - **complete**
 -   Fama-MacBeth regression - **complete**
@@ -128,21 +145,34 @@ regression. Currently only the single variable IV estimators are polished.
 
 ## Requirements
 
+**linearmodels 7.1 is the last release that supports Python 3.10 and NumPy 1.x
+(NumPy < 2).** Later releases will raise the minimum supported versions of
+Python, NumPy and the other dependencies.
+
 ### Running
 
 -   Python 3.10+
--   NumPy (1.22+)
+-   NumPy (1.22.3+)
 -   SciPy (1.8+)
 -   pandas (1.4+)
 -   statsmodels (0.13.1+)
--   formulaic (1.0.0+)
+-   formulaic (1.2.1+)
+-   pyhdfe (0.1+)
+-   mypy_extensions (0.4+)
+-   packaging
 -   xarray (0.16+, optional)
--   Cython (3.0.10+, optional)
 
+### Building from source
+
+-   A C compiler
+-   Cython (3.1.0+)
+-   meson-python
+-   NumPy (2.0+)
+-   setuptools_scm (9.2+)
 
 ### Testing
 
--   py.test
+-   pytest (9.0+)
 
 ### Documentation
 

SECURITY.md

@@ -0,0 +1,44 @@
+# Security Policy
+
+## Supported versions
+
+Only the latest release of linearmodels is supported. This is currently the
+7.1.x series. Security fixes are not backported to earlier releases, so please
+upgrade before reporting a problem that you found in an older release.
+
+## Reporting a vulnerability
+
+Please report suspected security vulnerabilities **privately**. Do not report
+them in a public issue, pull request or discussion, since that tells everyone
+about the problem before a fix is available.
+
+The preferred way to report a vulnerability is GitHub private vulnerability
+reporting:
+
+<https://github.com/bashtage/linearmodels/security/advisories/new>
+
+If you cannot use GitHub, email the maintainer, Kevin Sheppard, at
+<kevin.k.sheppard@gmail.com>, which is the address listed in `pyproject.toml`.
+Use a subject line that makes clear that this is a security report, for example
+"linearmodels security vulnerability".
+
+## What to include
+
+A report is easiest to act on if it contains:
+
+- A description of the vulnerability and its impact: what an attacker can do,
+  and what they need in order to do it.
+- The affected function, class or module, and the version of linearmodels where
+  you found the problem.
+- Steps to reproduce the problem, ideally as a short, self-contained script.
+- The versions of Python, NumPy, SciPy, pandas and statsmodels, and your
+  operating system.
+- A suggested fix or mitigation, if you have one.
+- Whether the problem has been disclosed elsewhere, any disclosure deadline that
+  you plan to follow, and whether and how you would like to be credited.
+
+## What to expect
+
+linearmodels is a volunteer project, so responses are best effort. There is no
+guaranteed response time or fix date. For a confirmed vulnerability, the
+maintainer will work with you on a fix and on the timing of its disclosure.

meson.build

@@ -11,8 +11,8 @@ cc = meson.get_compiler('c')
 cy = meson.get_compiler('cython')
 cython = find_program(cy.cmd_array()[0])
 
-if not cy.version().version_compare('>=3.0.10')
-  error('linearmodels requires Cython >= 3.0.10')
+if not cy.version().version_compare('>=3.1.0')
+  error('linearmodels requires Cython >= 3.1.0')
 endif
 
 py = import('python').find_installation(pure: false)

pyproject.toml

@@ -2,22 +2,23 @@
 build-backend = 'mesonpy'
 requires = [
   'meson-python',
-  "cython>=3.0.13,<4",  # Sync with CYTHON_MIN_VER in setup
+  "cython>=3.1.0,<4",  # Sync with the Cython check in meson.build
   "numpy>=2.0.0,<3",
   "setuptools_scm>=9.2.0,<10",
 ]
 
 [project]
 name = "linearmodels"
-# Keep syned with requirements.txt
+# Keep in sync with requirements.txt
 dependencies = [
   "numpy>=1.22.3,<3",
   "pandas>=1.4.0",
   "scipy>=1.8.0",
-  "statsmodels>=0.13.0",
+  "statsmodels>=0.13.1",  # The first release with a wheel for Python 3.10
   "mypy_extensions>=0.4",
   "pyhdfe>=0.1",
   "formulaic>=1.2.1",
+  "packaging",
 ]
 authors = [{ name = "Kevin Sheppard", email = "kevin.k.sheppard@gmail.com" }]
 classifiers = [
@@ -29,6 +30,8 @@ classifiers = [
   "Programming Language :: Python :: 3.11",
   "Programming Language :: Python :: 3.12",
   "Programming Language :: Python :: 3.13",
+  "Programming Language :: Python :: 3.14",
+  "Programming Language :: Python :: 3.15",
   "Operating System :: MacOS :: MacOS X",
   "Operating System :: Microsoft :: Windows",
   "Operating System :: POSIX",
@@ -37,6 +40,7 @@ classifiers = [
   "Topic :: Scientific/Engineering",
 ]
 description = "Linear Panel, Instrumental Variable, Asset Pricing, and System Regression models for Python"
+# The version is set in meson.build, which runs linearmodels/_build/git_version.py
 dynamic = [
   'version'
 ]
@@ -66,62 +70,56 @@ requires-python = ">=3.10"
 [project.optional-dependencies]
 test = [
   # test
-  "pytest>=8.4.1,<9",
+  "pytest>=9,<10",
   "pytest-randomly",
   "pytest-xdist",
   "pytest-cov",
   "coverage[toml]",
+  "xarray>=0.16",  # The tests import xarray
   "meson",
   "ninja",
 ]
 doc = [
   # Documentation
   "setuptools>=61",
   "wheel",
-  "setuptools_scm[toml]>=8.0.3,<9",
-  "oldest-supported-numpy",
-  "cython>=3.0.13,<4",  # Sync with CYTHON_MIN_VER in setup
+  "setuptools_scm[toml]>=9.2.0,<10",
+  "cython>=3.1.0,<4",  # Sync with the Cython check in meson.build
   "numpy >=1.22",
   "scipy >=1.5.0",
   "ipython >=8.0.1",
   "matplotlib >=3.0",
-  "patsy >=0.5.3",
   "statsmodels >=0.13.5",
   "jinja2",
   "sphinx>=7",
   "seaborn",
-  "numpydoc>=1.0.0",
+  "numpydoc>=1.10",
   "nbsphinx",
   "sphinx-immaterial",
   "jupyter",
   "notebook",
   "sphinx-autodoc-typehints",
   "fonttools>=4.43.0", # not directly required, pinned by Snyk to avoid a vulnerability
   "jupyterlab>=4.4.8", # not directly required, pinned by Snyk to avoid a vulnerability
-  "ipython>=7",
-  "sphinx_immaterial",
   "nbconvert",
   "pickleshare",
-  # Requirements for doc build, in addition to requirements.txt
-  "pandas~=2.3.3",
   "sphinxcontrib-spelling",
-  "sphinx_autodoc_typehints",
 ]
 dev = [
   "xarray>=0.16",
-  "pandas-stubsl",
+  "pandas-stubs",
   "wheel",
   # versioning
   "setuptools_scm[toml]>=9.2.0,<10",
   "packaging",
   # Performance
-  "cython>=3.0.10",
+  "cython>=3.1.0,<4",
   "numba>=0.55",
   # Graphics
   "matplotlib>=3",
   "seaborn",
   # Tests
-  "pytest>=8.4.1,<9",
+  "pytest>=9,<10",
   "pytest-randomly",
   "pytest-xdist",
   "pytest-cov",
@@ -134,6 +132,7 @@ dev = [
   "mypy>=1.3",
   "ruff>=0.16.0",
   "pyupgrade>=3.4.0",
+  "numpydoc>=1.10",  # The exclude_files setting of numpydoc lint needs 1.10
   "jupyterlab-code-formatter",
   "jupyterlab>=4.4.8",  # not directly required, pinned by Snyk to avoid a vulnerability
   "zipp>=3.19.1", # not directly required, pinned by Snyk to avoid a vulnerability
@@ -149,9 +148,6 @@ changelog = "https://bashtage.github.io/linearmodels/changes.html"
 [tool.meson-python.args]
 setup = ['--vsenv']
 
-[tool.setuptools_scm]
-version_file = "arch/_version.py"
-
 [tool.black]
 target-version = ['py310', 'py311', 'py312', 'py313', 'py314']
 exclude = '''
@@ -270,10 +266,10 @@ ignore = [
 "linearmodels/iv/results.py" = ["PERF203", "PLC0415"]
 
 [tool.pytest.ini_options]
-minversion = "8.4.1"
+minversion = "9.0"
 testpaths = "linearmodels"
 xfail_strict = true
-addopts = "--strict"
+strict = true
 empty_parameter_set_mark = "xfail"
 filterwarnings = [
   "ignore:`formatargspec`:DeprecationWarning:statsmodels",
@@ -289,7 +285,6 @@ filterwarnings = [
   "error:Support for multi-dimensional indexing:DeprecationWarning",
   "error:multivariate_ls is deprecated:FutureWarning",
   "error:invalid value encountered in true_divide:RuntimeWarning",
-  "ignore:The --strict option is deprecated, use --strict-markers instead:pytest.PytestDeprecationWarning",
   "ignore:`np.bool` is a deprecated alias:DeprecationWarning:pyhdfe",
   "ignore:defusedxml.cElementTree:DeprecationWarning:nbconvert",
   "ignore:the imp module is deprecated in favour of importlib:DeprecationWarning:pywintypes",
@@ -389,17 +384,19 @@ disallow_untyped_defs=true
 disallow_incomplete_defs=true
 show_error_context=true
 
-[mypy-linearmodels.conftest]
+[[tool.mypy.overrides]]
+module = "linearmodels.conftest"
 check_untyped_defs=false
 disallow_untyped_defs=false
 
-[mypy-linearmodels._version]
+[[tool.mypy.overrides]]
+module = "linearmodels._version"
 check_untyped_defs=false
 disallow_untyped_defs=false
 disallow_incomplete_defs=false
 
 [[tool.mypy.overrides]]
-module = "mypy-linearmodels.tests."
+module = "linearmodels.tests.*"
 check_untyped_defs=false
 disallow_untyped_defs=false
 disallow_incomplete_defs=false

requirements-dev.txt

@@ -2,15 +2,15 @@
 meson-python>=0.18.0
 meson
 ninja
-setuptools_scm[toml]>=8.0.3,<9
+setuptools_scm[toml]>=9.2.0,<10
 packaging
 wheel
 
 # Performance
-cython>=3.0.10
+cython>=3.1.0,<4
 
 # Testing
-pytest>=8.4.1,<9
+pytest>=9,<10
 pytest-xdist
 pytest-cov
 pytest-randomly
@@ -35,7 +35,7 @@ jupyterlab-code-formatter
 nbconvert
 nbformat
 nbsphinx
-numpydoc
+numpydoc>=1.10
 seaborn
 sphinx
 sphinx-immaterial

requirements.txt

@@ -1,8 +1,8 @@
 numpy>=1.22.3,<3
 pandas>=1.4.0
 scipy>=1.8.0
-statsmodels>=0.13.0
+statsmodels>=0.13.1
 mypy_extensions>=0.4
 pyhdfe>=0.1
-formulaic>=1.0.2
-wrapt!=2.0.0rc3
+formulaic>=1.2.1
+packaging